Rank 1 on App Store Top Free is not the same as leading Google Search. Muse from Meta, ChatGPT, Claude, and Gemini print different numbers on each tool.
Live data as of 2026-09-24
Muse from Meta is rank 1 on App Store Top Free while Google Search for muse ai still sits below ChatGPT. On September 24, 2026, get_top_trends type App Store Top Free stamped 2026-09-24T12:01:38 UTC listed Muse from Meta at 1, ChatGPT at 2, Google Gemini at 7, and Claude by Anthropic at 12. Dedicated get_growth on source google search, windows 7D, 30D, and 1Y, closed 2026-09-19 at muse ai 26, chatgpt 65, claude 62, and google gemini 14. Store rank and search index are different series. The search keyword is muse ai, not the chart label and not the single word muse.
The durable split is the one that survives next month's chart shuffle: iOS free-chart rank can put Muse from Meta above ChatGPT while Google Search still has ChatGPT more than twice as high. The App Store pull used limit 20, offset 0, count 20. Rank 3 was Vinted: Pre-loved marketplace. Rank 15 was WhatsApp Messenger. Grok AI was not in that top 20 on the rank sort.
| Chart label | App Store rank | Search keyword | Search index | 7D | 30D | 1Y |
|---|---|---|---|---|---|---|
| Muse from Meta | 1 | muse ai | 26 | -39.53% from 43 | +333.33% from 6 | +550.0% from 4 |
| ChatGPT | 2 | chatgpt | 65 | -23.53% from 85 | -17.72% from 79 | -35.0% from 100 |
| Google Gemini | 7 | google gemini | 14 | -12.5% from 16 | -6.67% from 15 | -86.0% from 100 |
| Claude by Anthropic | 12 | claude | 62 | -12.68% from 71 | -10.14% from 69 | +342.86% from 14 |
Every search row used the same recent_date 2026-09-19, the same 261 weekly points, and the same three windows. Muse's 7D fade is steeper than ChatGPT's, from a much smaller peak. ChatGPT fell from 85 to 65. Muse fell from 43 to 26. Claude is the closer search peer to ChatGPT at 62, even though Claude by Anthropic sits at App Store rank 12. Gemini is the inverse: App Store rank 7 next to search 14 and a 1Y drop of 86.0% from 100.
The 1Y baselines are 2025-09-20. Muse's year-ago index was 4. ChatGPT's was 100. Claude's was 14. Gemini's was 100. A launch that reprints App Store rank 1 can still be a mid-20s search series. Do not convert rank 1 into "most searched."
Meta's product post on 2026-09-08 describes Muse as a personal agent on Muse Spark, rolling out on iOS, Android, and muse.ai. That post does not publish a Google Search index. Trends MCP does.
Each AI name is its own get_growth request on source google search. Extra percent_growth windows in that call still count as one request per source plus keyword. The four rows above were four dedicated calls, not a comma list. chatgpt is lowercase. google gemini is two words. claude is one word. muse ai is two words.
{"mode": "get_top_trends", "type": "App Store Top Free", "limit": 20}
{"mode": "get_growth", "source": "google search", "keyword": "muse ai", "percent_growth": ["7D", "30D", "1Y"]}
{"mode": "get_growth", "source": "google search", "keyword": "chatgpt", "percent_growth": ["7D", "30D", "1Y"]}
{"mode": "get_growth", "source": "google search", "keyword": "claude", "percent_growth": ["7D", "30D", "1Y"]}
{"mode": "get_growth", "source": "google search", "keyword": "google gemini", "percent_growth": ["7D", "30D", "1Y"]}
MCP names match: get_top_trends(type="App Store Top Free", limit=20) and get_growth(keyword="muse ai", source="google search", percent_growth=["7D", "30D", "1Y"]). Omit percent_growth on muse ai and the handler defaults to 12M, which on this pull reprinted the 1Y row: 26 from 4, +550.0%, baseline 2025-09-20, 261 points. That default is safe on a five-year weekly search series. It is not safe on app rankings for a two-week-old Android package.
Bare muse is the trap. Source google search, same windows, closed 47 on 2026-09-19 from 54 on 2026-09-12, -12.96% 7D, +17.5% 30D from 40, +51.61% 1Y from 31. Estimated volume printed 815,000. That is not Meta's agent. muse from meta closed 23 from 31, -25.81% 7D, then 30D and 1Y printed 999999.0 from baseline 0.0, with data_quality low and a warning that the underlying signal is over 90% zeros. Store muse ai for search. Store Muse from Meta only as the App Store display name.
get_top_trends with type muse ai returned HTTP 400, error invalid_request, message Unknown type 'muse ai', plus 24 valid feed labels that include App Store Top Free. A keyword is not a board type. The live chart page is the App Store Top Free API. Window math lives on get-growth.
sort=rank_change with window=7d on App Store Top Free stamped the same as_of_ts 2026-09-24T12:01:38 UTC, plus baseline_ts 2026-09-17T02:01:38.593365 UTC, then returned count 25 climbers. Muse from Meta is absent. ChatGPT is absent. Google Gemini is absent. Claude by Anthropic is absent. The current number one cannot climb.
The 7d mover list opened at Capital One Shopping: Save Now, rank 17, prev_rank 73, rank_change 56. Netflix Game Controller sat at 45 from 80, rank_change 35. StoryReel: Drama Shorts & TV sat at 5 from 26, rank_change 21. Grok AI appeared at rank 31 from 37, rank_change 6. Facebook sat at 36 from 50, rank_change 15. A rank-change sort answers who moved into the chart, not who is already sitting at 1.
Default sort is rank. Omit sort to read Muse from Meta at the top. Keep rank_change for the climbers. Mixing the two sorts without quoting sort and baseline_ts will make Muse look like it vanished.
The Google Play board cannot patch this. Type Google Play, limit 25, still stamped as_of_ts 2026-09-13T12:01:38.463213 UTC on September 24, 2026. Rank 1 was WhatsApp Messenger. Rank 11 was ChatGPT. Rank 21 was Google Gemini. There was no Muse row in that 25. Play listing for bundle com.facebook.aura is titled Muse from Meta, 4.7 stars, 16.5K reviews, updated 2026-09-21, developer Meta Platforms, Inc. The frozen chart and the listing page are not the same object. Quote the Play as_of_ts if that board is used at all.
Android rank history is a third object. Source app rankings, keyword com.facebook.aura, windows 7D, 14D, 30D, closed index 100.0 and volume 1.0 on 2026-09-23, 12 points, data_start 2026-09-12. 7D and 14D printed growth 999999.0 from baseline index 0.0, while volume improved from 2.0 to 1.0, volume_growth -50.0%. 30D returned date_out_of_range. ChatGPT's bundle com.openai.chatgpt on the same source closed 100.0 / 1.0 across 7D, 14D, and 30D, 30 points from 2026-08-25, growth 0.0. Gemini's bundle com.google.android.apps.bard closed 92.0 / 6.0, 7D flat, 14D index +21.05% from 76.0 as volume improved from 10.0 to 6.0. Index 100.0 on two different packages is not proof they share a store slot.
REST is POST https://api.trendsmcp.ai/api with a Bearer key. MCP in Cursor is url https://api.trendsmcp.ai/mcp plus transport http. Connecting is free. Calling is not. Free is 100 requests per month. Starter is $19 for 1,000. get_top_trends on MCP costs 1 credit per 10 requested rows, max 10 credits. REST counts 1 request per call. get_growth bills once per source plus keyword.
A proving set that stays small: one App Store Top Free page, four google-search growth rows, one optional app rankings row on com.facebook.aura with 7D and 14D only. That is six requests if the four search names are each a call. Looping "until Muse looks bigger than ChatGPT" will spend the month reprinting the same 26-versus-65 split.
Search versus YouTube on that same phrase is a different page: Muse Google Search vs YouTube. The copy-paste tracker, including the Cursor JSON and the 400s, is how to track Muse from Meta.
Do not paste Muse from Meta into get_growth. Do not paste muse ai into get_top_trends as a type. The first is a chart label. The second is a search keyword. ChatGPT happens to use a chart label that is close to its search keyword. Muse does not.
FAQ